SuperSheet

Text to Columns in Excel: And Three Better Tools

By Mark Fulton · 2026-09-02 · 13 min read

Text to Columns in Excel: And Three Better Tools

Select the column, go to Data > Text to Columns, choose Delimited, tick the delimiter your data actually uses, and pick a Destination cell before you click Finish. That last step is the one everybody skips, and it is the reason the feature has a bad reputation: left alone, the wizard writes the split back over the original column and across as many columns to the right as the split produces, silently replacing whatever was sitting there. It also runs once. It remembers nothing, so when the same export lands next month you do the whole thing again by hand. For a one-off paste that is fine. For a column you will receive again, Flash Fill, TEXTSPLIT or Power Query are the better answers, and Text to Columns keeps one job the others cannot touch: forcing numbers and dates that arrived as text into real values.

Every walkthrough of this feature stops at Finish. Finish is the easy part.

What does the wizard actually do to your sheet?

Three screens, and each one makes a decision you can get wrong.

Microsoft's own steps are short: select the cell or column containing the text, select Data > Text to Columns, choose Delimited and Next, select your delimiters and preview the data, select Next, choose a Destination, then select Finish. The page lists it as applying to Excel for Microsoft 365, Excel 2024, Excel 2021, Excel 2019 and Excel 2016, and the path is the same on Windows and Mac desktop. Checked against that page on 2 September 2026.

What the three screens are really asking:

Screen one, Delimited or Fixed width. Delimited means "there is a character between the fields". Fixed width means "every field starts at the same character position", which is true of old mainframe and terminal reports and almost nothing else. If your preview shows ragged text, you want Delimited.

Screen two, the delimiters. You can tick more than one at a time, and there is a Treat consecutive delimiters as one checkbox that decides whether Smith,,Portland becomes two fields or three. The Text qualifier dropdown handles the case where a comma sits inside a quoted field, which is the difference between a clean split and a mess on any CSV that contains addresses or company names.

Screen three, Column data format and Destination. This is where the damage happens, and it is also where the feature earns its keep. More on both below.

Why does splitting overwrite the columns to the right?

Because the wizard writes into the worksheet, not into a new object, and its default Destination is the cell you started from.

A split that produces four fields needs four columns. It takes the one you selected plus the next three, and it does not ask whether anything is already living in them. There is no confirmation prompt when the target cells are empty, and the prompt you do get when they are not is easy to click through at speed. Excel's help pages for this wizard carry no warning about it at all, which is fair enough as documentation and unhelpful as a heads-up.

The fix is in the third screen and takes two seconds. Set Destination to the first empty column to the right of your data, $H$2 rather than $A$2, and the original column survives next to the result. Do that every time. It costs nothing when you were going to delete the original anyway, and it saves the file on the day the split turns out to be wrong.

There is a second, quieter version of the same problem. Text to Columns is a one-shot transformation with no record of itself. Nothing in the workbook says which column produced which, what delimiter was used, or that a split happened at all. Undo covers you until you save and close. After that the only trace is your memory.

This is also why the operation belongs early in a cleanup rather than late. Splitting a column that still carries trailing spaces just distributes the trailing spaces into four columns, and the whitespace is now in four places instead of one. The wider pre-send version of that ordering lives in the messy spreadsheet checklist.

When is Flash Fill the better tool?

When the pattern is obvious to a human and hard to describe as a delimiter.

Flash Fill watches what you type in a column beside your data and offers to finish the rest. Microsoft documents it at Data > Flash Fill, with Ctrl+E as the shortcut on Windows, and lists it as available in Excel for Microsoft 365, Excel 2024, 2021, 2019 and 2016, on Windows and Mac. Verified on 2 September 2026.

It is genuinely better than the wizard in three situations:

  • The split is not on a delimiter. Pulling 97214 out of Portland, OR 97214 is a pattern, not a character position. Type the first one, press Ctrl+E, done.
  • You want part of the string, not all of it. Text to Columns is all-or-nothing across the row. Flash Fill happily extracts one field and ignores the rest.
  • The formatting varies. Mixed capitalisation, an occasional missing middle name, a stray period. Flash Fill infers from two or three examples where a formula would need nested IFs.

Its weakness is the same as its strength. It guesses, and it does not tell you where it was unsure. The output is static values, so if row 400 was inferred wrongly you will not find out from a recalculation. Give it two or three examples rather than one, then spot-check the bottom of the column and any row that looked unusual in the source.

When does the split need to be repeatable?

The moment you can answer yes to "will this file arrive again".

That is the real dividing line, and it is not about how complicated the split is. A trivial split you repeat monthly is worse value than a fiddly split you do once. Two tools handle the repeat case.

TEXTSPLIT for a live formula. The syntax, per Microsoft, is =TEXTSPLIT(text, col_delimiter, [row_delimiter], [ignore_empty], [match_mode], [pad_with]). Two optional arguments matter for messy data: ignore_empty set to TRUE collapses consecutive delimiters, and pad_with replaces the #N/A that fills the gap when one row has fewer fields than the widest row. Version caveat, and it is a big one: the function's applies-to list on 2 September 2026 reads Excel for Microsoft 365, Excel for Microsoft 365 for Mac, Excel 2024 and Excel 2024 for Mac. Not Excel 2021, not 2019, and Excel for the web is not on that list either. If your recipients are on a perpetual licence, a TEXTSPLIT formula reaches them as an error.

Power Query when the file is an import. In the Power Query Editor the path is Home > Split Column > By Delimiter. The dropdown offers Colon, Comma, Equals Sign, Semicolon, Space, Tab or Custom, and the Split at options are left-most delimiter, right-most delimiter, or each occurrence of the delimiter. New columns are named after the original with a period and a number appended. The step is saved in the query, so next month's file is a refresh rather than a rebuild.

One messy address column, split four ways

Here is the column. Three rows, and the third field count is deliberately ragged, because real address exports always are.

A2: 1420 Cedar St, Apt 3B, Portland, OR 97214
A3: 88 Harborview Ave, Salem, OR 97302
A4: 5 Nettle Ln, Unit 12, Eugene, OR 97401

Row 3 has three comma-separated fields. Rows 2 and 4 have four. Every method below handles that differently, and that difference is the whole decision.

Method What lands in the sheet Where the ragged row goes Cost when the file arrives again
Text to Columns (Data > Text to Columns, Delimited, Comma) Static values across four columns, starting in the cell you selected Salem lands in the Apt column and OR 97302 lands in the City column. The whole row is shifted one column left Full manual repeat. Three screens, and you must re-set the Destination or overwrite again
Flash Fill (Ctrl+E, one column at a time) Static values, one column per pattern you demonstrate Usually right, because it reads the shape of each string rather than counting commas. It will not flag the row it guessed hardest on Full manual repeat, once per output column
TEXTSPLIT (=TEXTSPLIT(A2, ", ", , , , "")) A live spilled range, four cells wide, recalculating from A2 Fourth cell pads with the empty string you passed to pad_with, so nothing shifts. Values stay under the right heading Free. Paste new values into column A and the split follows. Needs Microsoft 365 or Excel 2024
Power Query (Home > Split Column > By Delimiter, each occurrence) A loaded table, refreshable, source untouched Blank in the missing field, columns stay aligned, and you can set the column count explicitly in Advanced options Free. Click Refresh. Highest setup cost, roughly five minutes, paid once

Read the third column of that table again. Text to Columns is the only method that silently misaligns the short row, and it misaligns it in a way that looks fine at a glance because every cell still contains plausible text. That is the failure mode worth designing around.

The advice underneath all of this is older than any of these features. Data Carpentry's spreadsheet lesson puts it plainly: don't include more than one piece of information in a cell, because it limits the ways you can analyse your data. Splitting is the repair. If you own the form or the export that produced the column, fixing it upstream beats repairing it monthly.

How does Text to Columns fix numbers and dates stored as text?

This is the job the wizard keeps, and it is the reason to leave it in your toolkit even after you have moved your real splits to Power Query.

Run Text to Columns on a single column with no delimiter that appears in the data, and click through to screen three. The Column data format options there are General, Text, Date, and Do not import column (skip). Microsoft's wizard reference spells out what each one buys you: to convert a column of all currency number characters to the Excel Currency format, select General, and to convert a column of date characters where each date runs year, month, day, select Date and then the YMD date type in the Date box. Pick the format, click Finish, and Excel re-parses every cell in the column as if it were being imported fresh.

That is a genuine conversion, not a formatting change. The values become real numbers and real dates, they align right, and SUM starts returning something other than zero.

Two things worth knowing. First, the same reference notes that Excel falls back to General when a conversion could produce unintended results, for instance when a column mixes alphabetical and numeric characters, or when the date order you chose does not match the data. So a column that stubbornly stays text after this usually has a row in it that is not what you think it is. Second, Microsoft's dedicated page on converting text-stored numbers does not mention this route at all. It covers the green error alert next to the cell and the VALUE function in a helper column. Both work. Neither handles a whole column of dates in one pass.

What is the Google Sheets equivalent?

Two of them, and they behave differently from each other.

The menu command is Data > Split text to columns, and Google's help page describes exactly that path, with a Separator dropdown that appears after the split so you can change the character Sheets guessed at. Same one-shot nature as Excel's wizard, same overwrite behaviour to the right, and Google's page carries no warning about that either.

The formula is SPLIT, with syntax SPLIT(text, delimiter, [split_by_each], [remove_empty_text]). There is one trap in it that catches people migrating from Excel. Google documents that by default each character in the delimiter is considered individually, so a delimiter of "the" splits the text around t, h and e separately. Passing ", " as a delimiter therefore splits on commas and on spaces, which shreds a street address. Set split_by_each to FALSE to treat the delimiter as one unit, or just split on the comma alone and trim afterwards.

SPLIT is live, so it fills the same role in Sheets that TEXTSPLIT fills in current Excel, and it is available to everyone rather than gated behind a version.

Frequently asked questions

How do I split a column without overwriting data?

Set the Destination on the third screen of the wizard to the first empty column to the right of your data instead of accepting the default, which is the cell you started from. Alternatively, insert as many blank columns as the split will produce before you run it, or use a formula method, since TEXTSPLIT in Excel and SPLIT in Sheets both write into a new range and leave the source column alone. If you have already run the split and saved the file, undo will not help you. Reopen the last version and start again.

What is the difference between Text to Columns and Flash Fill?

Text to Columns splits on a character you name and writes every resulting field across the row. Flash Fill infers a pattern from examples you type and fills one column at a time. Use the wizard when there is a real delimiter and you want all the pieces. Use Flash Fill when you want one piece, or when the boundary is a pattern rather than a character. Both produce static values, so neither updates when the source data changes.

Can I split by more than one delimiter?

Yes, in all four tools, with different limits. The Excel wizard lets you tick several delimiter checkboxes at once on screen two, plus a custom character in the Other box. TEXTSPLIT accepts an array as its col_delimiter, so {",", ";"} splits on either. Power Query's Split Column by Delimiter takes one delimiter per step, but steps stack, so you split twice. Google's SPLIT splits on each character of the delimiter string by default, which makes multi-delimiter splitting its default rather than an option.

Does Google Sheets have Text to Columns?

Yes, at Data > Split text to columns, verified against Google's help documentation on 2 September 2026. It auto-detects a separator and shows a dropdown afterwards so you can change it to comma, semicolon, period, space or a custom character. It overwrites columns to the right in the same way Excel's wizard does, so leave empty columns beside your data before you run it.

Fix the values before you split them

Splitting a column is the easy half. The half that decides whether the result is usable is what the values were already carrying: leading spaces that end up distributed across four columns, and numbers or dates that Excel is holding as text and will keep holding as text on the other side of the split.

That is what SuperSheet's spreadsheet cleaner reports on. Drop an .xlsx or .csv on it and you get a plan before anything changes, including the columns that contain numbers Excel is treating as text, named individually, along with a count of cells carrying stray spaces and the headers that would be re-cased. You approve the list or you close the tab, and either way you now know what the file is hiding.

Two honest limits, because the point of the plan is that it is accurate. The whitespace figure is a total across the sheet rather than a per-column breakdown, so it tells you how much there is rather than exactly where. And the cleaner does not split columns for you, the same way it deliberately does not unmerge cells. Which delimiter is the right one, and what each resulting column should be called, is a judgement about your data, and a tool that guesses at it produces a file you then have to check twice.

Run your file through it and see which columns come back flagged. Nothing uploads, the parsing and the download both happen in your browser, and the free tier does the whole job. Pro is $6 a month, billed $36 every six months, or $99 once, and it removes the export credit line, opens all five themes and raises the AI chat limits.


SuperSheet is a free set of spreadsheet formatting, CSV conversion, and cleanup tools. Your file never leaves the browser.